Torbjorn Zetterlund

Mon 07 2022

#datapipeline: Deploy Serverless Data pipeline to GCP with Terraform

by bernt & torsten

I recently built a data pipeline which I wrote up in an article data pipeline the hourly electricity rate, what I did not explain was how I deployed it to the Google Cloud Platform. In this article, I will show you how I used Terraform to deploy it to the Google Cloud Platform.

The biggest communication problem is we don’t listen to understand we listen to respond

Terraform is an open-source infrastructure as a code software tool that provides a consistent CLI workflow to manage hundreds of cloud services.

This post provides an overview of the Terraform resources required to configure the infrastructure for creating the Electricity Data Pipeline on Google Cloud Platform (GCP) as illustrated in the diagram below.

Electricity Data Pipeline
Electricity Data Pipeline

Terraform Resources

In order to define the above infrastructure in Terraform, we need the following Terraform resources:

Define the Provider

I am using Google I set my provider to Google and also define the GCP project name

provider "google" {
  project = "<project id>"
  region  = "us-central1"

Define a bucket

Next, I define the bucket to store the source code for the Cloud Function

resource "google_storage_bucket" "bucket" {
  name = "swedish-electricty-prices" # This bucket name must be unique
  location = "us"

Define Archive File

On my local machine, I define where my src is and where to store the file – Directory where your Python source code is

data "archive_file" "src" {
  type        = "zip"
  source_dir  = "${path.root}/../src" # 
  output_path = "${path.root}/../generated/"
resource "google_storage_bucket_object" "archive" {
  name   = "${data.archive_file.src.output_md5}.zip"
  bucket =
  source = "${path.root}/../generated/"

Define the Cloud Function

I define the Cloud Function resource, what region, what memory, what type of runtime and the entry point of the code. And from where to deploy the code to the cloud function.

resource "google_cloudfunctions_function" "function" {
  name        = "swedish-electricty-prices"
  description = "Trigger for Swedish Electricity Prices"
  runtime     = "python37"

  environment_variables = {
    PROJECT_NAME = "<Project Name",

  available_memory_mb   = 256
  timeout               = 360
  source_archive_bucket =
  source_archive_object =
  trigger_http          = true
  entry_point           = "main" # This is the name of the function that will be executed in your Python code

Define Service Account

I need to define a service account that can be accessed by the various GCP service automatically without my intervention.

resource "google_service_account" "service_account" {
  account_id   = "swedish-electricty-prices"
  display_name = "Cloud Function Swedish Electricty Prices Invoker Service Account"

Set IAM Role

For the Cloud Scheduler to be able to invoke a Cloud Function, I need to set a few IAM roles.

resource "google_cloudfunctions_function_iam_member" "invoker" {
  project        = google_cloudfunctions_function.function.project
  region         = google_cloudfunctions_function.function.region
  cloud_function =

  role   = "roles/cloudfunctions.invoker"
  member = "serviceAccount:${}"

Define the Cloud Scheduler

To wake up the Cloud Function I need to be able to create a scheduled job, I do that by using the Cloud Scheduler and in terraform define the details of setting a Cloud Schedule job.

resource "google_cloud_scheduler_job" "job" {
  name             = "swedish-electricty-prices"
  description      = "Trigger the ${} Cloud Function Daily at 6:00."
  schedule         = "10 0 * * *" # Daily at 12:10
  time_zone        = "Europe/Stockholm"
  attempt_deadline = "320s"

  http_target {
    http_method = "GET"
    uri         = google_cloudfunctions_function.function.https_trigger_url

    oidc_token {
      service_account_email =

There are many different ways to deploy a data pipeline using a flexible solution that makes full use of the capabilities of Terraform while creating a low-cost, scalable infrastructure. If you have any questions or ideas on how to improve this approach, you can make a comment below.